The bacterial cell wall is a critical load-bearing structure, but is not thought to be an important permeability barrier since proteins freely diffuse through isolated cell wall sacculi and bacteria secrete proteins without the aid of any known channels or transporters in the wall. Using new genetically encoded probes to measure the permeability of the cell in situ at the single-cell level, we discovered that the size threshold determining whether proteins can pass through the Bacillus subtilis sacculus is smaller than was previously thought. We found that transport of small proteins (\<10 kDa) through the sacculus was consistent with passive diffusion through discrete pores, while larger proteins (\>15 kDa) required the generation of larger pores by inducing peptidoglycan hydrolysis unbalanced by synthesis. These data are consistent with physics-based models of diffusion through a random percolation network of finite thickness. Conversely, the ability of the innate immune factor phospholipase (15.2 kDa) to kill B. subtilis was inhibited by membrane de-polarization. The protective effect of de-polarization was dependent on latent peptidoglycan synthesis (decoupled from cell growth) by PBP1 - highlighting a new role for this enzyme - and on reduced peptidoglycan hydrolysis. These results demonstrate that the rapid peptidoglycan turnover that drives cell growth also promotes movement of phospholipase across the cell wall, identifying a quintessentially bacterial mechanism of active transport.Competing Interest StatementThe authors have declared no competing interest.